Use the installed Council skill to review this Workshop Run Sheet app. First resolve available models and verify these exact provider/model routes: claudeAgent / claude-opus-5-5; codex / gpt-6.1-sol; opencode / openrouter/google/gemini-3.1-pro-preview. Report a missing route rather than substitute another provider. Use high effort where supported.

Maximum two review rounds, one implementation pass and 20 minutes total. Stop at that bound regardless of score. Reviewers are read-only with no nested delegates; the parent alone edits. No publishing, commits, pushes, purchases or changes outside this project.

Run node test.js first and fix known functional issues before review. Capture the app at 1440x1000 and 390x844, using default data: start10:00, target90, durations10/25/40/15. Give all three reviewers code and actual screenshots. Review hierarchy, legibility, contrast, spacing, locking/Fit/Undo clarity and mobile/keyboard usability. Preserve the existing Fieldnotes visual style and all calculations, preview/apply/undo, saved data, add/remove/reorder and CSV behaviour.

Collect concise specific findings and subjective scores, implement only justified improvements, run tests, then capture matched after screenshots and request the second review from all three providers. Report unresolved issues, disagreement and inaccurate observations honestly. Save screenshots, exact model identities, a diff and a final report. Cost is unknown unless usage is actually available. Stop and close the temporary preview server when finished.
